Bot crashing due to the error "aiohttp.http_websocket.WebSocketError: Received fragmented control frame"

[nttstar]

Dear authors, I've encountered this error today, and it leads to the entire bot process crashing. Is there a solution or a way to avoid this?
(version 2.3.2)

Full stack:

File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/discord/client.py", line 860, in run
    asyncio.run(runner())
  File "/opt/conda/envs/pytorch/lib/python3.9/asyncio/runners.py", line 44, in run
    return loop.run_until_complete(main)
  File "/opt/conda/envs/pytorch/lib/python3.9/asyncio/base_events.py", line 647, in run_until_complete
    return future.result()
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/discord/client.py", line 849, in runner
    await self.start(token, reconnect=reconnect)
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/discord/client.py", line 778, in start
    await self.connect(reconnect=reconnect)
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/discord/shard.py", line 468, in connect
    raise item.error
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/discord/shard.py", line 164, in worker
    await self.ws.poll_event()
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/discord/gateway.py", line 626, in poll_event
    raise msg.data
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/client_ws.py", line 196, in close
    msg = await self._reader.read()
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/streams.py", line 657, in read
    return await super().read()
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/streams.py", line 627, in read
    raise self._exception
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/client_ws.py", line 229, in receive
    msg = await self._reader.read()
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/streams.py", line 657, in read
    return await super().read()
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/streams.py", line 616, in read
    await self._waiter
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/http_websocket.py", line 288, in feed_data
    return self._feed_data(data)
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/http_websocket.py", line 295, in _feed_data
    for fin, opcode, payload, compressed in self.parse_frame(data):
  File "/opt/conda/envs/pytorch/lib/python3.9/site-packages/aiohttp/http_websocket.py", line 459, in parse_frame
    raise WebSocketError(
aiohttp.http_websocket.WebSocketError: Received fragmented control frame

[Rapptz]

This is a protocol level error so it's not really fixable on my side.

[nttstar]

Is it possible to catch all runtime WebSocketError?

[Rapptz]

That does not seem wise

[nttstar]

@Rapptz Hello author, this error occurred twice again last night. I'm wondering if it could be caused by network instability on the server during some time period? Would upgrading the aiohttp library potentially fix it（currently it's 3.8.3)? Or is there some way we could hook some code to ignore exceptions like this that cause the whole program to exit?

[issamansur]

Hi. Can u please give the part of code where this error throwing? Or what is your bot doing before crash?

[nttstar]

@issamansur https://github.com/Rapptz/discord.py/blob/master/discord/gateway.py#L628 . As a temporary workaround, I commented out this line of code to see if it helps.

[issamansur]

It's source of discord.py)
I asked about the actions of the bot when the program stopped working. Apparently, this is a "unique" error. I think that maybe you are using some barely used functionality, where there may be some problems with data transmission.
P.S. well, or just packets "break" during transmission due to a connection failure)))

[nttstar]

@issamansur I just used the standard features of dpy, and it works well in most cases. This error occurs very rarely. I believe it could be a possible WebSocket network error, possibly due to temporary hardware or network issues. In theory, dpy should handle or ignore such errors.

[Rapptz]

This will be fixed in v2.4 (163a86c)

[nttstar]

You are super!

[nttstar]

@Rapptz May I ask when the 2.4 version is expected to be released?

[Rapptz]

No ETA, sorry.

[llimonix]

an error consistently occurs once a week